Universidad de Extremadura
CRITERIOS DE EVALUACIÓN
A manuscript to be submitted to the Atmospheric Environment
Abstract
In recent years the corn grain ethanol industry has expanded and led to increased availability of dried distillers grains with solubles (DDGS), and feeding DDGS to swine is becoming more common in pork production. With feed being the primary cost in pork production and increasing interest in air emissions from animal feeding operations, it is important to understand the impacts of non-traditional dietary formulations on aerial emissions. The purpose of this study was to evaluate
the impacts of feeding DDGS on ammonia (NH3), hydrogen sulfide (H2S) and
greenhouse gas (GHG) emissions from deep-pit swine wean-to-finish (5.5 – 118 kg) facilities in Iowa, the leading swine producing state in the USA. To attain the study objectives, two commercial, co-located wean-to-finish barns were monitored: one barn received a traditional corn-soybean meal diet (designated as Non-DDGS regimen), while the other received a diet that included 22% DDGS (designated as DDGS regimen). Gaseous concentrations and barn ventilation rate (VR) were monitored or determined semi-continuously, and the corresponding emission rates
(ER) were derived from the concentration and VR data. Two turns of production were monitored for this study, covering the period of December 2009 to January 2011. The daily and cumulative emissions are expressed on the basis of per barn, per pig, and per animal unit (AU, 500 kg live body weight). Results from this project indicate that feeding 22% DDGS does not significantly affect aerial emissions of
NH3, H2S, CO2, N2O or CH4 when compared to the Non-DDGS regimen in a deep-pit
wean-to-finish swine facility (p-value = 0.10 for NH3, 0.13 for H2S, 0.55 for CO2, 0.58
for N2O, and 0.18 for CH4). ER for the Non-DDGS regimen, in g/d-pig, averaged 7.5
NH3, 0.37 H2S, 2127 CO2 and 72 CH4. In comparison, ER for the DDGS regimen, in
g/d-pig, averaged 8.1 NH3, 0.4 H2S, 1849 CO2, and 48 CH4. On the basis of kg gas
emission per AU marketed, the values were 8.7 NH3, 0.724 H2S, 2350 CO2 and 84
CH4 for the Non-DDGS regimen; and 12 NH3, 0.777 H2S, 2095 CO2, and 60 CH4 for
the DDGS regimen. Results of this extended field-scale study help filling the knowledge gap of GHG emissions from modern swine production systems. Keywords: Ammonia, Hydrogen sulfide, Greenhouse gases, Emissions, Swine
Introduction
Iowa leads the United States in corn and ethanol production. For corn-based ethanol plants, a primary co-product of the process is distillers dried grains with solubles (DDGS). DDGS have been reported to contain high levels of digestible energy and metabolizable energy, digestible amino acids, and available phosphorus (Shurson et al., 2003; Honeyman et al., 2007). Generally, DDGS have been found to contain 2 to 3.5 times more amino acids, fat, and minerals than corn (Honeyman et
al., 2007). Animal nutritionists have suggested including up to 20% DDGS in nursery, grow-finish, and lactating sow diets and up to 40% in gestating sows and boars (Honeyman et al., 2007). However, the decision to feed DDGS is generally based on economics. At the current DDGS and corn prices the inclusion of DDGS in swine diets has provided a cost savings over traditional non-DDGS diets.
It has been hypothesized that sulfur levels in DDGS could result in increased
hydrogen sulfide (H2S) emissions from stored swine manure when pigs are fed
rations containing DDGS. However, comparative data from full-scale swine production systems are needed to confirm any impacts on air emissions from feeding DDGS. The increased usage of DDGS at swine facilities has led several researchers to examine the effect of DDGS on emissions, odors, and manure composition, but these studies have been at lab or at non-commercial scales and the data from these studies were inconsistent (Spiehs et al., 2000; Gralapp et al., 2002; Xu et al., 2005; Jarret et al., 2011)
Spiehs et al. (2000) performed a 10-week trial on 20 barrows receiving either a DDGS (at a 20% inclusion rate) or non-DDGS ration. The pigs were housed, based on diet, in two fully-slatted pens within the grow-finish room of a swine research facility. The non-DDGS diet was a typical corn-soybean meal; total phosphorus and total lysine were held constant in both diets within each phase of
feeding. The study was conducted to evaluate differences in odor, H2S, and
ammonia (NH3)from stored manure as a result of the pig’s diet. The stored manure
that was evaluated for emissions was maintained in a container to simulate deep-pit storage. Air samples were collected from the headspace of storage containers. Over
the 10-week period, this study reported that DDGS (at a 20% inclusion level) did not
affect odor, H2S, or NH3 emissions from the stored manure.
Gralapp et al. (2002) performed six, four week trials utilizing a total of 72 finishing pigs. Three diets containing 0, 5, 10% DDGS were fed during the study. Manure from the study was collected in a pit below each environmental chamber where the pigs were housed. Samples were collected on day 4 and day 7 of each week and analyzed. Each pit was cleaned weekly. Gralapp et al. (2002) observed no significant differences between concentrations of total solids (TS), volatile solids (VS), chemical oxygen demand (COD), total kjeldahl nitrogen (TKN), and
phosphorus (TP) content. Additionally, this study compared the effects on odor of each of the different diets and found there were no significant differences.
Xu et al. (2005) performed a study utilizing 40 nursery pigs to evaluate
phosphorus excretion from animals receiving DDGS diets. The diets contained 0, 10, 20% DDGS. Results indicated that diets containing 10 and 20 % DDGS had a 15 and 30 % increase in daily manure excretion, respectively, compared to pigs fed the corn-soybean meal diet. Xu et al. (2005) reported the increase was due to a 2.2 and 5.1 % reduction in dry matter digestibility in rations containing 10 and 20 % DDGS, respectively. Reportedly, reduced dry matter digestibility was the result of increased amounts of crude protein and higher fiber levels in the DDGS diet.
Jarret et al. (2011) investigated the effects of different biofuel co-products (DDGS, SBP, and high fat level rapeseed meal on nitrogen (N) and carbon (C) excretion patterns as well as ammonia and methane emissions. Ammonia emissions
ammonia traps. Biochemical methane potentials (BMPs were then ran on the manure to determine the methane production potential of the difference diet
regimens. The DDGS diet was found to excrete the more N, C and dry matter than the other rations. It was also reported that diets with higher fiber contents with higher crude protein (CP) inclusions were had similar ammonia emissions as lower fiber and lower protein diets. Methane production potential was also found to be the lowest in manure when pigs were fed DDGS.
The results from these studies cannot be directly compared because of differences in rations, animal housing, manure storage, and analytical methods. Besides differences in the experimental design of these studies, the results may also be affected by scaling issues. Additionally, only two of the studies investigated the effects of feeding DDGS to swine on aerial emissions, both were small scale
experimental studies. This has led to deficit of data concerning the impact of DDGS on air emissions at the farm scale.
The primary objective of this study was to quantify the impact on gaseous emissions of feeding DDGS to wean-to-finish pigs in two commercial deep-pit swine barns. The secondary objective was to compare the emission results of this study to similar full-scale emission monitoring studies that have been reported in the
literature. To meet these objectives, NH3, H2S and greenhouse gases (GHG)
(carbon dioxide – CO2, nitrous oxide – N2O, and methane – CH4) concentrations
were measured and emission data were collected using a mobile air emissions monitoring unit (MAEMU). The results were further compared with the available literature data.
Methods and Materials Site Description
Two 12.5 x 57 m (50 x 190 ft) co-located wean-to-finish deep-pit swine barns, designated as Non-DDGS and DDGS, located in central Iowa were monitored for two production turns. Pigs entered the barns at 5.5 kg (12 lbs) and were marketed at 118 kg (260 lbs). Each turn was approximately 27 weeks in length with pigs entering the barns at 3 weeks and marketed around 30 weeks of age. The barns had a rated capacity of 1,200 marketed pigs. Both barns were double-stocked initially, meaning during the wean-to-grow (W-G) phase (first 6 to 10 weeks of the turn) both barns held approximately 2,400 pigs. When the pigs reached 27 kg (60 lbs), approximately half of the pigs were moved off-site to another facility for the grow-to-finish (G-F) phase. Each barn had four 0.6 m (24 in.) pit fans, two 0.6 m (24 in.) endwall fans for mechanical ventilation, and sidewall curtains on both sides to provide natural
ventilation when needed. The barns were equipped with three space heaters 66 kW (225,000 BTU/h) each, 20 brooder heaters 5 kW (17,000 BTU/h) each and 20 bi-flow ceiling inlets (one per pen).
The diets used during this study were formulated to meet the pigs’
requirements as they grew towards market weight (NRC, 1998); the only difference in ingredients between the Non-DDGS (control) diet and the DDGS (treatment) diet was the inclusion of 22% DDGS for the DDGS regimen. The ingredients and diet formulations used during this study are proprietary information. Including DDGS resulted in higher levels of crude protein, crude fiber, acid detergent fiber and sulfur compared to the non-DDGS diet. The nursery phase diets for both barns did not
include DDGS. Nursery diets were fed until the pigs weighed 12 kg or approximately 10 to 14 days after entering the barn. Therefore, data for the periods when nursery diets were fed were excluded from the analysis.
The producer provided weekly pig performance data, including mortality and
average body weight for the duration of the project.
Measurement System
A MAEMU was used to continuously collect emissions data from the two deep-pit wean-to-finish swine barns. The instruments and data acquisition system were housed in the MAEMU. A detailed description of the MAEMU and operation can be found in Moody et al. (2008). Constituents measured during this study were
NH3, CO2, N2O, CH4, and H2S. Aerial emissions were monitored for two growout
periods. A photoacoustic multi-gas analyzer (INNOVA Model 1412, INNOVA AirTech
Instruments A/S, Ballerup Denmark) was used to measure NH3, CO2, N2O, and CH4
concentrations. H2S concentrations were measured using an ultraviolet fluorescence
H2S analyzer (Model 101E, Teledyne API, San Diego, CA). The instruments were
challenged weekly with calibration gases and recalibrated as needed. All calibration gases were certified grade with ± 2% accuracy.
Air samples were drawn from three composite locations (north pit fans, south pit fans, and endwall fans) in each barn and an outside location to provide ambient background data (Figure 1). Each composite sampling location was chosen to match the fan stages used at the facility. Pit fan sampling points were located below the slats next to each fan. Endwall sample ports were placed approximately 1.0 m (3.28
ft) in front of each endwall fan. Sample locations and placement of sampling ports were chosen to ensure representativeness of the air leaving the barns. Air samples were collected in 30-s cycles for four cycle periods (120 s) at each location. The fourth reading from each sampling cycle was used as the measured pollutant concentration. Use of the fourth reading was due to the fact that the INNOVA and API had T98 and T95 response time of 120 s and 100 s, respectively. Each
sampling point had three consecutive dust filters (60, 20, 5 µm) to keep particulate
matter from plugging or contaminating the sample lines, the servo valves, or the delicate instruments.
A positive-pressure gas sampling system (P-P GSS) was used in the MAEMU to prevent introduction of unwanted air into the sampling line. The P-P GSS
consecutively pumped sample air from each sampling location using individual designated pumps. Air samples from each location were collected sequentially over the 120 s period via the controlled operation of servo valves of the PP-GSS. Each barn sampling location was sampled every 14 min. It was assumed with the
sequential sampling that any concentration change at a given location between two sampling periods followed a linear relationship. Therefore, linear interpolation was used between sampling points to determine the intermediate concentrations and to line up the concentration with the continuously measured ventilation rate (VR) for the location. A background ambient air sample was collected every two hours for 8 minutes. Background concentrations were subtracted from the exhaust readings when air emissions rates were calculated for the barns. All pumps and the gas
sampling system were leak checked weekly to ensure no contamination was occurring.
Pit fans at this facility had variable speeds, while the endwall fans had a single speed. All fans were calibrated in situ at multiple operation points (RPM and static pressure) to develop a performance or airflow curves for each fan. The in situ calibration of the exhaust fans was conducted with a fan assessment numeration system (FANS) (Gates et al. 2004). For single-speed fans (endwall), airflow was a function of static pressure, whereas for variable-speed fans, airflow was a function of static pressure and fan speed (revolution per minute or RPM). Runtime of each fan
was monitored continuously using an inductive current switch(with analog output)
attached to the power cord of each fan motor (Muhlbauer et al., 2011). Each current switch’s analog output was connected to the data acquisition (DAQ) system
(Compact Fieldpoint, National Instruments, Austin, Tex) (Li et al., 2006). Both barns were equipped with static pressure sensors (model 264, Setra, Boxborough, Mass.). Each pit fan’s RPM was continuously measured using Hall Effect speed sensors (GS100701, Cherry Corp, Pleasant Prairie, WI). Atmospheric pressure, indoor and outdoor temperature, and relative humidity (RH) were measured with barometric pressure sensor (WE100,Global Water, Gold River, Cal.), temperature sensors (type-T thermocouple, Cole Palmer, Vernon Hills, Ill.), and RH probes (HMW60, Vaisala, Woburn, Mass.). Signals were sampled every second and averaged and recorded on the on-site computer in 30 second intervals.
VR during periods of natural ventilation was determined using a CO2 balance,
principal of indirect animal calorimetry (Xin et al., 2009). More specifically, the
metabolic heat production of non-ruminants is related to oxygen (O2) consumption
and CO2 production of the animals (Brouwer, 1965) ( Equation 1). Using this
relationship the VR can be estimated by using the inlet and exhaust CO2
concentrations and the total heat production (THP) of the animals (Equations 2 & 3). For the purpose of this study, finishing pig THP under thermoneutrality (Pedersen and Sallvik, 2002) (Equation 4) and a respiratory quotient (RQ) of 1.14 was used.
` (1) Where, THP = total heat production rate of the animals (W)
O2 = oxygen consumption rate of the animals (mL s-1)
CO2 = carbon dioxide production rate of the animals (mL s-1)
(2)
` (3)
Where, VR = building ventilation rate (m3 s-1)
CO2 = carbon dioxide production rate of the animals (mL s-1)
CO2 e = carbon dioxide concentration of exhaust (ppmv)
CO2 i = carbon dioxide concentration of inlet (ppmv)
2 2
5.02*
*
18
.
16
O
CO
THP=
+
2 2 2,
02
.
5
/
18
.
16
O
CO
RQ
where
RQ
THP
CO
=
+
=
i eCO
CO
CO
VR
2 2 2−
=
(4) Where, THP = total heat production rate of animals (W)
m = mass of animal (kg)
n = daily feed energy intake (expressed as n times the maintenance requirement)
Body mass used in the THP calculation was provided weekly from the
producer and linearly interpolated for daily values. The daily feed energy intake was calculated using information provided by the producer about feed composition and the daily maintenance energy requirement (DME, kcal/day) for a finishing swine provided by NRC (1998) (Equation 5). Calculated values for n ranged from 6.9 to 2.9 (with an average of 3.5) for pig weights from 5 -120 kg, respectively.
(5)
Where, BW = animal body weight (kg)
In addition to air sampling, manure samples were collected monthly from each barn. Manure samples were collected from each of the four pit pump-out locations and composited for each barn. Samples were cooled and shipped to Midwest Laboratories (Omaha, NE) and were analyzed for total solids (TS), total
nitrogen (TN), ammoniacal nitrogen (NH3-N), total phosphorus (TP), potassium (K),
sulfur (S), calcium (Ca), magnesium (Mg), sodium (Na), iron (Fe), manganese (Mn),
]
09
.
5
09
.
5
*
)][
003
.
47
.
0
(
1
[
09
.
5
m
.75m
n
m
.75m
.75THP=
+
−
+
−
75 . 0*
106 BW
DME=
copper (Cu), zinc (Z), and pH. A total of eleven manure samples from each barn was collected and analyzed during the monitoring period.